Abstract: In order to investigate deeply the application status of on-line chemical instruments in China, onsite dynamic inspection and analysis were performed on 695 chemical instruments of 11 power plants across 9 provinces using YHJ-V inspection device. The error variance diagram was plotted based on the big data from tests, and then both the current situation and existing problems are summarized. It is concluded that the calculation method recommended by DL/T 246—2015《Chemical Supervision Guidelines》is difficult to be implemented to evaluate the measurement accuracy and supervision reliability of on-line chemical instruments. To be more specific, there is no better way to accurately count the number of total hours and hours of normal operations due to lack of technologies to verify the instrumental measurements. Therefore, a new model for evaluating the accuracy of chemical instruments in electric power industry has been proposed in this paper, which can more objectively calibrate the accuracies of the instrumental measurement and quantitatively determine the difference of instrument management and maintenance levels, which is beneficial to motivate the maintenance team in their routine work so as to take full advantages of chemical instruments for supervision purposes.

Key words: power plant, chemical instruments, on-line inspection, dynamic test, error dispersion, evaluation model